Adsorption of a Nonionic Symmetric Triblock Copolymer on Surfaces with Different Hydrophobicity

Transcription

1 pubs.acs.org/langmuir 2010 American Chemical Society Adsorption of a Nonionic Symmetric Triblock Copolymer on Surfaces with Different Hydrophobicity Xiaomeng Liu, Dong Wu, Salomon Turgman-Cohen, Jan Genzer, Thomas W. Theyson, and Orlando J. Rojas*,,^, Department of Forest Biomaterials, Department of Chemical and Biomolecular Engineering, and Department of Textile Engineering, Chemistry, and Science, North Carolina State University, Raleigh, Raleigh, North Carolina 27695, Department of Chemistry, Duke University, Durham, North Carolina 27708, and ^Faculty of Chemistry and Materials Sciences, Department of Forest Products Technology, Aalto University, P.O. Box 16300, 00076, Aalto, Finland Received January 12, Revised Manuscript Received March 14, 2010 ) ) ) This study investigates the adsorption of a symmetric triblock nonionic polymer comprising ethylene oxide (EO) and propylene oxide (PO) blocks (Pluronic P-105, EO 37 PO 56 EO 37 ) on a range of substrates including hydrophobic, i.e., polypropylene (PP), poly(ethylene terephthalate) (PET), nylon, and graphite, and hydrophilic, i.e., cellulose and silica. The adsorption process and the structure of the hydrated adsorbed layers are followed by quartz crystal microgravimetry (QCM), surface plasmon resonance (SPR), and atomic force microscopy. The unhydrated surfaces are characterized by ellipsometry and contact angle techniques. The adsorption kinetics and the extent of adsorption are determined by monitoring the changes in resonance frequency and refractive index of sensors coated with ultrathin films of the various substrates. Langmuirian-type adsorption kinetics is observed in all cases studied. The amount of adsorbed Pluronic on hydrophobic polymer surfaces (PP, PET, and nylon) exceeds that on the hydrophilic cellulose. The hydrophobic (graphite) mineral surface adsorbs relatively low polymer mass, typical of a monolayer, while micellar structures are observed on the hydrophilic silica surface. The amount of water coupled to the adsorbed polymer layers is quantified by combining data from QCM, and SPR are found to increase with increasing polarity of the substrate. On the basis of contact angle data, the nonhydrated adsorbed structures produce modest increases in hydrophilicity of all the substrates investigated. Overall, insights are provided into the structure and stability of both hydrated and nonhydrated adsorbed triblock copolymer. Introduction Hydrosoluble polymers are often used to adjust the interfacial properties of materials by self-assembly to generate structured functional surfaces. As such, modification of solid surfaces by physical adsorption of triblock copolymers has received increased attention This is in part due to their amphiphilic properties that endows molecular constructs with tailorable surface affinity to the surrounding media. They can deliver steric stabilization in solid dispersions, generate controlled surface structures on a range of materials, and modify interfacial properties such as wetting and lubrication. 8 Recent interest has been prompted due to their potential as drug-delivery vehicles in aqueous solution where micelles contain a hydrophilic corona and a hydrophobic core within which drugs can be solubilized and transported. 12,13 *Corresponding author. address: ojrojas@ncsu.edu, Tel: þ , Fax: þ (1) Cosgrove, T. J. Chem. Soc., Faraday Trans 1990, 86, (2) van de Steeg, L. M. A.; Golander, C. G. Colloids Surf. 1991, 55, 105. (3) Malmsten, M.; Linse, P.; Cosgrove, T. Macromolecules 1992, 25, (4) Shar, J. A.; Obey, T. M.; Cosgrove, T. Colloids Surf. 1999, 150, (5) Brandani, P.; Stroeve, P. Macromolecules 2004, 37, (6) Shar, J. A.; Obey, T. M.; Cosgrove, T. Colloids Surf. 1998, 136, (7) Brandani, P.; Stroeve, P. Macromolecules 2003, 36, (8) Napper, D. H. Polymeric Stabilization of Colloidal Dispersions; Academic Press: London, (9) Linse, P. Macromolecules 1993, 26(17), (10) Nolan, S. L.; Philips, R. J.; Cotts, P. M.; Dungan, S. R. J. Colloid Interface Sci. 1997, 191, (11) Alexandridis, P.; Holzwarth, J. F.; Hatton, T. A. Macromolecules 1994, 27, (12) Foster, B.; Cosgrove, T.; Hammouda, B. Langmuir 2009, 25, (13) Foster, B.; Cosgrove, T.; Espidel, Y. Langmuir 2009, 25, The properties of triblock copolymers in aqueous solution 9-11 and their adsorptive behavior have been addressed by a number of authors. 1-7 The adsorption of EO-PO-EO block copolymers on mineral (silica, mica, graphite) and self-assembled hydrophilic surfaces, 3-5 polystyrene, and self-assembled hydrophobic surfaces 6,7,14 have been reported. The effect of the relative size of the constituent blocks on adsorption has been addressed also, but only few studies have considered the detailed structure of the adsorbed layer. Brandani and Stroeve, for example, characterized the structure of adsorbed triblock copolymer layers as highly dependent on the relative length of the hydrophilic and hydrophobic blocks of the macromoleucle. 7 Nevertheless, the relation between aqueous bulk solution and adsorbed, interfacial properties of triblock copolymers remains a largely unexplored subject, especially in the case of materials relevant to applications such as textiles and papermaking. In our previous work, 15 we described the manufacture of ultrathin, flat films of polypropylene (PP), polyethylene (PE), poly(ethylene terephthalate) (PET), nylon, and cellulose by spincoating. This was proposed as a simple platform for investigations involving interfacial phenomena by using techniques such as quartz crystal microgravimetry (QCM), ellipsometry, surface plasmon resonance (SPR), and reflectometry. In this study, similar substrates deposited on QCM and SPR sensors are employed to probe the adsorption from aqueous solution of a (14) Green, R. J.; Tasker, S.; Davies, J.; Davies, M. C.; Rojberts, C. J.; Tendler, S. J. B. Langmuir 1997, 13(24), (15) Song., J.; Liang, J.; Liu, X.; Krause, W. E.; Hinestroza, J. P.; Rojas, O. J. Thin Solid Films 2009, 517, Langmuir 2010, 26(12), Published on Web 03/31/2010 DOI: /la100156a 9565

2 Article typical symmetric Pluronic block copolymer. More specifically, the effect of aqueous polymer concentration on the extent and dynamics of adsorption and desorption is addressed. The viscoelasticity of adsorbed EO-PO-EO polymer and the amount of coupled water are considered. In addition, the structure of the adsorbed EO-PO-EO polymer is examined by using atomic force microscopy (AFM) operated in fluid medium. Finally, the change in water contact angle of the different substrates and the thickness and stability of the adsorbed EO-PO-EO layers are determined. Materials and Methods Deionized water from an ion-exchange system (Pureflow, Inc.) followed by treatment in a Milli-Q Gradient unit with a resultant resistivity of >18 MΩ 3 cm was used to prepare the polymer solutions and was employed as a medium in QCM, SPR, and AFM experiments (background fluid, rinsing solution, etc.). A symmetrical triblock nonionic copolymer consisting of ethylene oxide (EO) and propylene oxide (PO) blocks, under the trade name of Pluronic P105 (BASF Corporation) was used without further purification. This polymer comprised two hydrophilic poly-eo terminal blocks, each having an average of 37 EO units, and a central, relatively hydrophobic poly-po block having an average of 56 PO units (for a theoretical weight-average molecular weight of 6500 Da). Hereafter, the polymer is denoted as EO 37 PO 56 EO 37. The experimental molecular weight of EO 37 PO 56 EO 37 was determined by size-exclusion chromatography coupled with light scattering (Wyatt MiniDAWN) and differential refractive index detection (OptilabREX instruments.) The resulting number average molecular weight and polydispersity index were found to be 6300 Da and 1.04, respectively (see Figure 1 of Supporting Information and related discussion). These values were in agreement with determinations reported elsewhere from MALDI- TOF MS measurements. 12 The cloud point and solubility in aqueous solution of EO 37 PO 56 EO 37 were determined to be C and >10%, respectively. A relatively small micellar aggregation number of 2.9 was reported recently for aqueous solutions at 5% EO 37 PO 56 EO 37 concentration at 25 C. 12 Aqueous solutions having polymer concentrations ranging from to 10 w/v % in water were freshly prepared before each reported experiment. Silica wafers (cut into smaller pieces of cm 2 )were obtained from Wafer World Inc., FL. Graphite wafers were obtained from SPI Supplies Division of Structure Probe, Inc., PA. QCM gold-coated quartz sensors (Q-Sense Inc., Sweden) and SPR gold slides were used in QCM and SPR experiments, respectively. These sensors were treated first with Piranha solution (70% H 2 SO 4 þ 30% H 2 O 2 (30%)) for 20 min followed by UV/ozone treatment (28 mw/cm 2 at 254 nm wavelength) for 10 min to remove any organic contaminants. Thin films of PP, PET, nylon, and cellulose were deposited on the cleaned QCM and SPR gold sensors by spin-coating; details about their manufacture can be found in ref 15. Samples for water contact angle and thickness measurements, before and after treatment with EO 37 PO 56 EO 37 solutions, were fabricated on flat silica wafers that were subjected to same cleaning and spin-coating procedures used in QCM experiments with PP, PET, nylon, and cellulose. Surface Tension. The surface tension of EO 37 PO 56 EO 37 solutions was determined by using the Wilhelmy plate method with a Cahn balance (Thermo Material Characterization, USA, Madison, WI). Surface tension isotherms at 25 C were recorded and the critical micelle concentration (CMC) as well as the surface excess at the air-liquid interface was determined. All EO 37 - PO 56 EO 37 solution concentrations were thereafter reported as multiple units of the CMC (CMC = 381 μm; see later). Quartz Crystal Microbalance. The rate of EO 37 PO 56 EO 37 adsorption, the adsorbed mass, and characteristics of the adsorbed layer, described further below, were assessed using a quartz crystal microbalance (Q-Sense model E4, Gothenburg, Liu et al. Sweden). The changes of resonant frequency f and energy dissipation D of the polymer-coated QCM sensors were measured simultaneously by switching on and off the applied voltage. The shift in the resonance frequency was used to calculate the areal adsorption by means of the Sauerbrey equation (eq 1) 16 which is generally applicable if (1) the adsorbed macromolecules form a thin, rigid, and homogeneous layer, and (2) the extra mass deposited on the sensor is small compared to that of the resonator (polymer-coated sensor). Δm ¼ cδf ð1þ n In eq 1, c represents a constant characteristic of the sensitivity of the resonator to changes in mass (17.7 ng Hz -1 cm -2 for the used 5 MHz quartz crystals) and n is the overtone number (n =1,3,5, 7, etc.). QCM senses the adsorbed mass with no distinction between contributions from the adsorbed polymer mass and its coupled or hydration water. However, since SPR is less sensitive to the degree of hydration, the water associated with the adsorbed molecules was estimated by comparing the adsorbed mass obtained from SPR (Δm SPR )andqcm(δm QCM ) experiments, according to the following expression: 17 %coupled water ¼ Δm QCM - Δm SPR Δm QCM! 100 The change in QCM energy dissipation D wasusedtodetermine the viscoelastic properties of the adsorbed layer. D was typically measured after switching off the resonator and by recording the exponential decay in oscillation (frequency and amplitude dampening), which were then used to obtain the energy dissipated and stored during one period of oscillation, E dissipated and E stored, respectively, according to eq 3 D ¼ E dissipated ð3þ 2πE stored Energy dissipation can be attributed to (1) changes in the viscoelastic properties of the crystal and adsorbed layer and (2) variations in the density and viscosity of the surrounding solution. 18 Therefore, changes in f and D were recorded after a rinsing step to replace the adsorbing EO 37 PO 56 EO 37 solution with pure water, thereby allowing comparison of f and D values before and after adsorption so as to obtain effective changes in these parameters. The QCM measurement modules and tubing were cleaned for one hour before each run by using a 2% (v/v) Hellmanex solution (Hellma GMBH, M ullheim, Germany). They were then rinsed with ethanol and water. After mounting the respective polymer-coated sensor in the QCM module, water was injected continuously with the system adjusted to a temperature of ( 0.02 C. In a typical experiment, the different polyolefin, polyester, and cellulose uniform films were first deposited on the QCM gold sensors by spin-coating. The thicknesses and roughness of the respective thin films under same operating conditions were reported in our previous publication. 15 The shifts in QCM frequency, both in air and in water, were used to test the quality of the coating before each experiment. Prior to any measurement, the polymer-coated sensors were allowed to equilibrate in water for half a day in order to establish the base f and D signals, which were then zeroed. In order to study the dynamics of EO 37 PO 56 EO 37 adsorption, aqueous solutions of EO 37 PO 56 EO 37 were injected into the QCM (16) Sauerbrey, G. Z. Phys. 1959, 155, 206. (17) Hedin, J.; Lofroth, J.; Nyden, M. Langmuir 2007, 23, (18) Reimhult, E.; Larsson, C.; Kasemo, B.; Hook, F. Anal. Chem. 2004, 76, ð2þ 9566 DOI: /la100156a Langmuir 2010, 26(12),

3 Liu et al. flow module at a constant flow rate of 0.1 ml/min. The shifts in f and D were followed as a function of time for 25 min, followed by rinsing with pure water. QCM adsorption data were obtained by running single experiments at the given polymer concentration; however, experiments with sequential increase in concentration of the injected polymer solutions were also carried out. All adsorption experiments were conducted at least in triplicate and average values reported. Surface Plasmon Resonance. Complementary information about EO 37 PO 56 EO 37 adsorption on the studied surfaces was obtained by surface plasmon resonance (SPR 200, KSV Instruments, Ltd., Finland). SPR measures variation in surface plasmons excited by external light source. The extreme sensitivity of electromagnetic plasmon waves, which propagate along the interface between a metallic substrate and the surrounding medium, to any variation at the interface is ideally suited for detecting molecular adsorption events. 12 Because the SPR signal or the changes in optical resonance properties define the specific angle shift, the SPR signal expressed in resonance units is a measure of the mass adsorbed on the sensor surface. 19,20 Therefore, the resonance units (or refractive index units, 1 RU = ) are easily converted into areal mass. The adsorbed mass per surface area is linearly proportional to ΔRU according to 18,21 Δm SPR ¼ C SPR ΔRU where the proportionality constant, C SPR, is different for different species. For EO 37 PO 56 EO 37, C SPR was determined from the change in refractive index with concentration, dn/dc, which amounted to 0.14 ng/cm 2. 7 The SPR experiments were performed under the same set of conditions as those used in QCM work (EO 37 PO 56 EO 37 solution concentration, module temperature, flow rate, rinsing protocol, etc.) so that complementary information on adsorption and desorption behaviors could be obtained. Atomic Force Microscopy. The morphologies of adsorbed EO 37 PO 56 EO 37 were investigated on model surfaces consisting of silica and graphite by a MFP3D system (Asylum Research, Inc., Santa Barbara, CA) operating in the tapping mode and in liquid conditions using the idrive technique. 22 The surfaces were chosen to represent the hydrophobic (PP, PET, nylon) and hydrophilic (cellulose) surfaces used in QCM and SPR experiments, which had a small yet measurable roughness that prevented detailed observation of the EO 37 PO 56 EO 37 structures formed. The AFM tips were Al backside, coated SiN with a force constant of 0.08 N/m and a resonance frequency in the range khz. The image set point was set to V and image analyses were performed offline using the software provided by the AFM manufacturer and WSxM (Nanotech Electronica, USA). A first scan of bare silica and graphite surfaces in pure water gave the references for each sample. Then, the AFM experiments consisted of passing the polymer solution, waiting approximately 1 h to achieve equilibrium conditions, and then performing a scan on the same sample. Scans were acquired on a wide area varying from nm 2 to μm 2. A planar background was subtracted from the data to compensate for tilt of the sample relative to the scanning plane. Ellipsometry. A single wavelength ellipsometer (Rudolf, Model Auto EL) was employed to measure the thicknesses of the polymeric substrates and to assess their change after adsorption of EO 37 PO 56 EO 37. The experimental procedure involved immersion of polymer-coated silica wafer into the EO 37 PO 56 EO 37 (19) Homola, J.; Yee, S. S.; Gauglitz, G. Sens. Actuators, B 1999, 54, (20) Liedberg, B.; Nylander, C.; Lundstrom, I. Biosens. Bioelectron. 1995, 10, R1 R9. (21) Rodahl, M.; Kasemo, B. Sens. Actuators, A 1996, 54, (22) AsylumResearch Data Sheet 24, idrive Magnetic Actuated Cantilever for Effortless Cantilever Tunes in Fluid. Langmuir 2010, 26(12), ð4þ Article Figure 1. Surface tension isotherm for aqueous solutions of EO 37 - PO 56 EO 37 measured at 25 C. The lines are added as guides to the eyes. aqueous solution overnight, at a concentration 4 times above the CMC, rinsing with water, and drying the system gently under a nitrogen jet before mounting it in the ellipsometer stage. In order to test the stability of the adsorbed EO 37 PO 56 EO 37 layer, the thicknesses of the surfaces after adsorption of the polymeric surfactant were also measured after the surfaces were stored for 4 h at 80 C. The refractive index, n, and the optical (ellipsometric) thickness, δ e, were obtained from the changes in the ellipsometric angles Δ and Ψ. Water Contact Angle. The water contact angles (WCA) on the different surfaces were measured with a Phoenix 300 system (SEO Corporation, Korea) by computer-controlled application of a water droplet (4 μl volume) from a syringe. The images of the sessile drop were analyzed with respect to their width and height to yield the contact angle and drop volume by using the ImageJ software (National Institutes of Health, USA). The contact angles of the polymer substrates were assessed both before and after EO 37 PO 56 EO 37 adsorption from 4 CMC aqueous solutions after the substrates were gently blown with nitrogen gas in a laminar flow cabinet. The averages of at least three contact angles for each condition and substrate are reported here. In order to check the stability of the EO 37 PO 56 EO 37 layers adsorbed on the polymer substrates, water contact angles after strong sonication in water (42 khz ( 6%) were also measured. Results and Discussion Figure 1 shows the surface tension isotherm of aqueous solutions of EO 37 PO 56 EO 37. The change in the surface tension slope indicated a critical micelle concentration (CMC) of 381 μm, in agreement with other reports. 9,11 It was proposed that at this low EO 37 PO 56 EO 37 concentration the molecules associated as hard-sphere micelles. 10 The slight reduction in surface tension after the CMC was indicative of the anticipated dispersity in molecular mass and surface activity of EO 37 PO 56 EO 37 in the sample. The surface excess Γ at the air/liquid interface was obtained by the Gibbs adsorption equation for nonionic surfactants Γ ¼ - 1 RT γ ln c where γ is the surface tension, in the present case measured at constant temperature of K, R is the universal gas constant, and c is the EO 37 PO 56 EO 37 molar concentration. At maximum polymer packing, the calculated surface excess was molecules/nm 2 (0.122 nm 2 per molecule or 878 ng/cm 2 ), which is T ð5þ DOI: /la100156a 9567

4 Article Liu et al. Figure 2. Mean values of third overtone QCM frequency (a) and dissipation (b) as a function of time for EO 37 PO 56 EO 37 adsorption on PP surfaces at various aqueous solution concentrations ( μm). The experiments were conducted in an open (continuous) flow configuration with EO 37 PO 56 EO 37 solution injection rate of 0.1 ml/min (starting at about 450 s). The dip observed in all profiles soon after the adsorption plateau (at ca s) was produced after rinsing the system with water. In (c), refractive index signals from same experiments with surface plasmon resonance are presented. Similar behaviors to those observed for PP were obtained in the case of PET, nylon, and cellulose surfaces (data not shown). noted here as a reference for further comparison with the areal adsorption determined at solid-liquid interfaces (see later sections). Adsorption of EO 37 PO 56 EO 37 on Polymeric Surfaces. QCM was used to study the adsorption of EO 37 PO 56 EO 37 from solutions with different polymer concentrations on thin films of PP, PET, nylon, and cellulose. In the case of the hydrophobic substrates, the triblock copolymer was expected to adsorb with the hydrophobic PO block bound or anchored on the surface and the EO blocks forming buoyant, swollen structures. QCM and SPR experiments were carried out in order to shed light on this hypothesis, to quantify the adsorption, and gain details about the conformation of the adsorbed layer. The plateau QCM third overtone frequency f and dissipation D values after EO 37 PO 56 EO 37 adsorption were recorded before and after rinsing with water. Figure 2 shows QCM data to reveal the dynamics of the adsorption process in the case of EO 37 - PO 56 EO 37 adsorbing on PP surfaces; similar profiles were obtained for the other polymeric substrates. In a typical adsorption experiment, water was first injected continuously in the QCM sample loop until stable f and D baselines were achieved. After injection of the respective EO 37 PO 56 EO 37 solution in the QCM module, which contained the respective coated sensor, sharp shifts in frequency and dissipation were observed (see Figure 2). These changes were indicative of a rapid mass uptake on the PPcoated resonator due to the adsorption of EO 37 PO 56 EO 37 molecules. After the frequency and dissipation signals reached a plateau value, typically in less than 10 min, rinsing water was injected to remove any excess or loosely bound EO 37 PO 56 EO 37. The recorded signals were used to measure the effective adsorption by comparing them with the frequency and dissipation baselines, in the absence of adsorbing polymer, under the same bulk solution density and viscosity. Upon rinsing, abrupt changes in frequency and dissipation were observed until reaching constant values, typically within 10 min. The reversible and kinetically irreversible adsorbed masses were calculated from the frequency signals before and after rinsing, i.e., first and second frequency plateau, respectively. The frequency shifts in QCM, and thus the adsorbed mass, increased with EO 37 PO 56 EO 37 solution concentration. The adsorption of EO 37 PO 56 EO 37 on PP surfaces was expected to occur with the PO segments anchoring on the surface while the EO groups were solvated in the aqueous medium. In fact, it has been reported that EO 37 PO 56 EO 37 adsorbed on highly hydrophobic surfaces under the so-called brush regime. 7 However, the details of the adsorbed layer could be very complex, especially if substrates with different hydrophobicity are considered. For example, a phase-segregated layer structure with interacting PO groups forming on the hydrophobic PP surface, via hydrophobic forces, has been reported. 11 Furthermore, it has been suggested that single molecular chains or micellar aggregates of block copolymers diffuse from solution to adsorb on unoccupied PP surface sites, followed by a slow buildup of surface density by the penetration of chains through the existing monolayer, combined with a molecular rearrangement leading to a brush conformation for the solvated EO groups. 14,23 Similar processes could be expected to occur in the case of the present triblock copolymer, which adsorbed with a larger density as the solution concentration was increased. However, at some limiting polymer concentration access to the unoccupied adsorption sites on the surface was hindered because molecular chains would need to penetrate through the segregated hydrophilic EO layers. This produced adsorption saturation, as was observed by the fact that at the highest concentration, 15.4 mm, the frequency shift after rinsing was smaller than that at lower concentration (0.15 and 1.54 mm). As was shown by Brandani and Stroeve, the reduction in adsorption (or in the present case the negative of frequency shift) observed at the highest concentrations could be due to a kinetically induced metastable equilibrium with progressively less efficient packing at the hydrophobic surface. 7 Changes in D, depicted in Figure 2b, suggest that before rinsing with water EO 37 PO 56 EO 37 adsorbed as a loose structure, while the adsorbed EO 37 PO 56 EO 37 remaining on the surface after rinsing showed the characteristic energy dissipation of a relatively (23) Johner, A.; Joanny, J. F. Macromolecules 1990, 23, DOI: /la100156a Langmuir 2010, 26(12),

5 Liu et al. thin and rigid structure (a net dissipation shift of no more than units). Supporting evidence of this hypothesis was obtained after comparing the normalized QCM frequency and dissipation for the various overtones (data not shown). In such cases, the adsorbed mass after rinsing could be accurately calculated with the Sauerbrey relation (eq 1). The changes in the refractive index due to EO 37 PO 56 EO 37 adsorption on the surfaces were monitored in complementary SPR experiments. Once a stable refractive index baseline was obtained in aqueous medium, EO 37 PO 56 EO 37 was injected and data collected until a plateau value was reached for each solution concentration investigated. Water was then used to rinse out any loosely adsorbed macromolecules. The mean values of SPR data after EO 37 PO 56 EO 37 adsorption on PP from different solution concentrations, before and after rinsing with water, are plotted in Figure 2c. As was the case in QCM experiments, the amount of adsorbed EO 37 PO 56 EO 37 was found to increase with increasing solution concentration. Both QCM and SPR experiments showed that at submicellar EO 37 PO 56 EO 37 concentrations the shift of frequency and refractive index were larger, while at high polymer concentrations (from 0.15 to 1.54 mm), the signals changed slowly until reaching values indicative of a maximum adsorption or adsorption saturation. After rinsing with buffer solution, the refractive index was smaller than before rinsing indicating that loosely adsorbed EO 37 PO 56 EO 37 molecules were removed leaving only the more tightly bound adsorbed ones on the PP surface. Since SPR is less sensitive to surrounding solution density and viscosity, these results support previous QCM data in that the change in frequency (or effective adsorbed mass) observed after rinsing was caused by adsorption of EO 37 PO 56 EO 37 in addition to any possibly minor change in the physical properties of the surrounding medium. The kinetics of PEO-PPO-PEO triblock copolymer adsorption onto a hydrophobic self-assembly surface was investigated by Brandani and Stroeve. 24 In their work, the initial adsorption rate prior to rinsing was found to depend on the structure of the adsorbing Pluronic polymer. A simple Langmuir kinetic model as that used by Karpovich and Blanchard 25 was found suitable to fit the adsorption data. This model considered the adsorbed mass per unit area, Γ, which is a function of time, t, the maximum adsorbed mass, Γ m, and the empirical kinetic constants K 0 and k obs (see eq 6 below and more details in Supporting Information). ΓðtÞ ¼ K 0 ½1 - expð - k obs tþš ð6þ Γ m Also, the initial adsorption rate was found to depend on the polymer aqueous solution concentration (see Supporting Information for QCM and SPR data). In Figure 3a,b, the QCM frequency isotherms of EO 37 - PO 56 EO 37 adsorbing on all surfaces studied (PP, PET, nylon, and cellulose) are presented, before and after rinsing with water. It can be observed that rinsing did not produce complete desorption of EO 37 PO 56 EO 37. In fact, frequency shifts after rinsing indicated the removal of loosely bound EO 37 PO 56 EO 37 molecules, but a large number of segments remained adsorbed on the surface, likely due to an energy barrier that prevented their desorption. The bulk of irreversible contacts are expected to be ascribed to the PO blocks but the existence of adsorbing EO segments cannot be ruled out. On the basis of the frequency shifts, it can be concluded (24) Brandani, P.; Stroeve, P. Macromolecules 2003, 36, (25) Karpovich, D. S.; Blanchard, G. J. Langmuir 1994, 10, Langmuir 2010, 26(12), Article that the extent of EO 37 PO 56 EO 37 adsorption increased with the substrate hydrophobicity; it was largest on PP, followed by PET, nylon, and cellulose. Adsorption on the different hydrophobic surfaces (PP, PET, and nylon) could be expected to occur in the form of monolayers. However, as was noted before, preadsorbed molecules may have prevented the formation of uniform layers; instead, aggregated surface structures may have formed with extensive chain penetration and brush-like, solvated EO groups protruding in solution. The WCA of bare PP, PET, and nylon surfaces were 102, 69, and 61, respectively; therefore, it is hypothesized that a larger hydrophobic driving force (related to the WCA) led to more effective adsorption of the EO 37 PO 56 EO 37 molecules and/or a stronger ability to overcome the steric barriers imposed by adsorbed chains that otherwise would have prevented further adsorption. In contrast to hydrophobic surfaces, adsorption on the hydrophilic cellulose surface resulted in a very small frequency shift. In fact, the frequency signal was only measurable after a threshold EO 37 PO 56 EO 37 concentration, equivalent to the CMC. Furthermore, the irreversible adsorption of EO 37 PO 56 EO 37 tended to reach a maximum in the frequency shifts near the vicinity of the CMC. These observations suggest that a structural change occurred at a concentration above the CMC, i.e., adsorption became evident as micelles formed in solution. Figure 3c,d shows the change in refractive index for all surfaces (PP, PET, nylon, and cellulose) before and after rinsing with water, respectively. The trends observed followed closely those noted in the QCM experiments. Specifically, adsorption of EO 37 PO 56 EO 37 onto hydrophobic surfaces resulted in larger refractivity shifts at the different equilibrium concentrations, while lower adsorption was observed in experiments with hydrophilic cellulose surfaces. For relatively hydrophobic surfaces (PP, PET, nylon), the shifts in refractive index after adsorption of EO 37 PO 56 EO 37 from submicellar aqueous solutions displayed a continuous change with increasing concentration. However, in the case of cellulose surfaces negligible adsorption was measured at submicellar concentrations. After rinsing with buffer solutions, similar tendencies were observed, while at high concentration ( 1-4 CMC), a maximum adsorption was reached. Figure 3b,d show the main features of the adsorption isotherms, with a maximum in the adsorbed amount observed at a concentration close to the CMC. Deviations to simple Langmuir isotherms have been observed in the case of adsorption on selfassembled monolayers of long chain alkanethiols 7 and in other related systems. 3,26 Such effects have been interpreted in terms of the polydispersity of the sample, with macromolecules of larger molecular mass displacing the smaller ones at high solution concentration due to the crossover in adsorption mechanism. 26,27 Water Coupled to Adsorbed EO 37 PO 56 EO 37 Layers. The complementary nature of QCM and SPR data facilitates insightful information about the adsorption of EO 37 PO 56 EO 37 on the different surfaces. The frequency shifts of QCM depended on the total oscillating mass, including water coupled to the adsorbed molecules, while for SPR, the refractivity was not affected by bound water molecules. 28 Hence, by calculating the adsorbed mass from QCM and SPR the contribution of water coupled with or solvating the adsorbed layer could be evaluated. (26) Munch, M. R.; Gast, A. P. J. Chem. Soc., Faraday Trans. 1990, 86, (27) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Springer: Berlin, (28) Hook, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Chemistry 2001, 73, DOI: /la100156a 9569

6 Article Liu et al. Figure 3. QCM frequency (a,b) and SPR refractive index (c,d) profiles for PP, PET, nylon, and cellulose upon adsorption of EO 37 PO 56 EO 37 before rinsing with water (reversible adsorption, a and c) and after rinsing with water (irreversible adsorption, b and d). The experimental standard deviations for all data collected are shown as error bars at each condition. The solid lines are added as a guide to the eyes. In order to compare the respective maxima of adsorbed masses, the shifts in frequency (QCM) and refractivity index (SPR) after rinsing were converted into adsorbed mass by using eqs 2 and 3, respectively. Table 1 summarizes the results of such calculation for all surfaces exposed to aqueous solutions of EO 37 PO 56 EO 37 at 4 CMC concentration, after rinsing with water. The trends in the adsorption isotherms obtained by using these two techniques were in very good agreement (see Figures 2 and 3). QCM s Sauerbrey adsorbed masses were higher than the respective SPR values. This difference was ascribed to the contribution of coupled water (trapped water, hydration and solvation water), associated with ethylene oxide (EO) groups. The amount of EO 37 PO 56 EO 37 adsorbed on the hydrophobic substrates (PP, PET, and nylon) calculated from QCM experiments after rinsing was 16-38% higher than the SPR values, depending on the surface. Such a difference was distinctively larger (76%) in the case of hydrophilic cellulose substrates (see Table 1). Similar values of percent coupled water for related systems have been reported from SANS experiments. 12 QCM and ellipsometric data have also been compared for adsorbed nonionic Table 1. EO 37 PO 56 PO 37 Adsorbed Mass (QCM and SPR) from 1.54 mm (4 CMC) Solution Concentration on Different Surfaces after Rinsing with Water QCM adsorbed mass, ng/cm 2 (molecules/nm 2 ) SPR adsorbed mass, ng/cm 2 (molecules/nm 2 ) % coupled water (eq 2) PP 249 (0.21) 210 (0.19) 16.0 PET 169 (0.11) 106 (0.10) 37.5 Nylon6 108 (0.08) 75 (0.07) 30.4 Cellulose 37 (0.03) 9 (0.01) 76.2 surfactants, and mass overestimations of 101% and 73% were found for hydrophilic and hydrophobic substrates, respectively. 29 Finally, comparison of the SPR adsorbed mass on the more hydrophobic PP substrate (0.19 molecules/nm 2 )withthatobtained from surface tension measurements revealed that adsorption on the solid surface occurred only to a low extent with respect to the adsorption at the air-liquid interface (maximum packing of 0.81 molecules/nm 2 ). Such observation can be ascribed to the (29) Sta lgren, J. J. R.; Eriksson, J.; Boschkova, K. J. Colloid Interface Sci. 2002, 253, DOI: /la100156a Langmuir 2010, 26(12),

7 Liu et al. Article Figure 4. ΔD-Δf profiles revealing changes in the conformation during adsorption of EO 37 PO 56 EO 37 from solution concentrations below the CMC (0.001%, 1.54 μm) (a) and above the CMC (1%, 1540 μm) on the different surfaces investigated. ΔD-Δf profiles shown in (c) and (d) correspond to the profiles shown in (b) (adsorption of EO 37 PO 56 EO 37 from solution concentrations above the CMC), but the data are separated for the cases before rinsing (c) and after rinsing (d). higher entropic penalties in the case of solid interfaces compared to the less dense air-water interface. 7 Conformation of EO 37 PO 56 EO 37 Adsorbed Layers. As evidenced by QCM frequency and SPR refractive index shifts (Δf and ΔRIU), EO 37 PO 56 EO 37 adsorption on different substrates exhibited distinctive differences in the extent of adsorption and the kinetics of the process. These differences were attributed to the mass and structure of the EO 37 PO 56 EO 37 adsorbed layers, the relationship of which can be better discussed in light of QCM s ΔD and Δf relationships. To this end, ΔD-Δf plots were employed, with the time variable explicitly eliminated and, as concluded in previous studies, with the absolute ΔD-Δf slope and respective slope variation providing information about associated kinetic regimes and conformational changes. 30 Figure 4 shows the QCM ΔD-Δf profiles for adsorption of EO 37 PO 56 EO 37 from submicellar (a) and higher (b-d) solution concentrations. The slopes of these curves for the different substrates indicated that EO 37 PO 56 EO 37 (30) Su, X. D.; Zong, Y.; Richter, R.; Knoll, W. J. Colloid Interface Sci. 2005, 287, Langmuir 2010, 26(12), tended to form a rigid adsorbed layer at submicellar concentrations (Figure 4a). Loops in the ΔD-Δf profile were observed, which were typical consequence of water rinsing. At high polymer concentration (above CMC, Figure 4b), the ΔD-Δf profiles during adsorption of EO 37 PO 56 EO 37 indicated a more dissipative adsorbed layer. The respective buildup of the layer at EO 37 - PO 56 EO 37 solution concentrations above the CMC and the changes after rinsing (irreversible adsorption) are shown more clearly in Figure 4c,d, respectively. The curves for EO 37 PO 56 EO 37 adsorption on PP exhibited a steep slope, indicative of a hydrated adsorbed layer, while adsorbed layers on the less hydrophobic surfaces were much less dissipative, in agreement with the previous findings of percent coupled water. A possible description of the buildup of soft adsorbed layers could include the initial formation of a thin, patchy layer followed by an increased adsorbed mass as more molecules diffused to the interface, even forming loosely bound multiple layers. Since the binding between these layers was expected to be weak, compared with the molecules in the close vicinity to the surface, they could be easily removed by rinsing (see loops in the ΔD-Δf profiles). Furthermore, the molecules that remained at the interface, after DOI: /la100156a 9571

8 Article Liu et al. Figure 5. AFM 1 μm scans of silica surfaces in DI water (a) and in at 0.4 CMC and 4 CMC EO37PO56EO37 aqueous solution concentrations (b and c, respectively). AFM 200 nm scans of silica surfaces in DI water (d) and in at 0.4 CMC and 4 CMC EO37PO56EO37 aqueous solution concentrations (e and f, respectively). Representative line section analyses in each case are also provided (g-i). rinsing, were expected to be bound more rigidly (see Figure 4d). Finally, clear differences in the ΔD-Δf data for the hydrophobic and hydrophilic substrates were observed in Figure 4, which highlight the differences in substrate hydrophobicity and thereby characteristic extent of adsorption. AFM imaging was employed to unveil more details about the morphology of adsorbed EO37PO56EO37 molecules on bare silica and graphite surfaces, below or above the CMC. These substrates were chosen as representative for the two distinctive types of surfaces employed in this investigation, i.e., hydrophobic and hydrophilic, and were also selected due to their lower roughness which could allow a clearer and direct visualization of detailed adsorbed features. In fact, the small but yet sizable and inherent roughness of spin coated films of PP, PET, nylon, and cellulose prevented such efforts, especially when imaging was performed in aqueous solution, as reported here. Figure 5 displays AFM images acquired in aqueous solutions of EO37PO56EO37 at submicellar concentrations and above the CMC. They show adsorption features on silica at two different scan sizes. Figure 6 corresponds to the case of adsorption on the graphite substrate. Figure 5a,d for bare silica (for two different size scans) immersed in pure water did not show any feature. Figure 5b,e, depicting the topographic images of the surface after exposure to EO37PO56EO37 solution of submicellar concentration, did not show major differences, as can be better observed in Figure 5g,h corresponding to the respective section analyses. Therefore, there was indication that a smooth adsorbed layer was formed on silica. In contrast, Figure 5c,f showed a distinct change in the surface morphology when the concentration of EO37PO56EO37 was above the CMC: globular surface aggregates, presumably 9572 DOI: /la100156a adsorbed EO37PO56EO37 micelles were detected. While the scan size in Figure 5c allowed an assessment of the surface density of the adsorbed micelles, Figure 5f corresponding to smaller, 200 nm, scan size is presented in order to determine more clearly the metrics of such features. Figure 5f,i indicated globular aggregates with diameter around nm and a vertical dimension of 1 nm. The shape and dimensions of the adsorbed micelles were in agreement to those reported for the respective micelles in aqueous solutions, i.e., a radius of gyration of 8.6 nm ( 20 nm diameter)10 except that the anchor PO block assumes a flatter configuration on the substrate. Similar observations have been reported for a copolymer with a higher molar number of the hydrophobic block (P103 or EO17PO60EO17) that produced association structures characterized by (AFM) vertical and lateral dimensions of 1 and 30 nm, respectively, after adsorption on a self-assembled hydrophobic surface.7 In contrast to the observations discussed in relation to silica surfaces in Figure 5, adsorption of EO37PO56EO37 on graphite did not produce any distinguishable feature, as shown in Figure 6. The changes in texture and roughness were rather small, regardless of the concentration of polymer used. Thus, it can be concluded that smooth and flat layers were formed after adsorption of EO37PO56EO37 on hydrophobic, graphite surfaces. Such a layer is expected to be the result of dangling EO chains protruding in solution. The smooth and compact layer observed at concentrations above the CMC agrees with results reported by Brandani and Stoeve for other surfaces.7 Changes in WCA and Integrity of the Adsorbed EO37PO56EO37 Layers. The layer thickness and water contact angle were measured after EO37PO56EO37 adsorption on the solid Langmuir 2010, 26(12),

9 Liu et al. Article Figure 6. AFM 200 nm scans of graphite surfaces in DI water (a) and in at 0.4 CMC and 4 CMC EO 37 PO 56 EO 37 aqueous solution concentrations (b and c, respectively). Representative line section analyses in each case are also provided (d-f). Table 2. Water Contact Angles and Thicknesses of the Surfaces before and after Adsorption of EO 37 PO 56 EO 37 under Different Drying Conditions bare surfaces surfaces after EO 37 PO 56 EO 37 adsorption drying at 25 C drying at 80 C thickness, (nm) WCA (deg) thickness (nm) WCA (deg) thickness (nm) WCA (deg) PP PET nylon cellulose surfaces from solutions above the CMC (4 CMC) (see Table 2). In each case, adsorption was allowed overnight followed by water rinsing and drying with a gentle nitrogen jet. Additional samples were dried after adsorption at 80 C under vacuum. The water contact angle on bare PP surfaces was the highest, 102, while those for PET, nylon, and cellulose were 69, 61, and28, respectively. The highest WCA for PP is explained by the inherently higher hydrophobicity of PP since it contains only -CH 2 - and -CH 3 groups, while nylon and PET contain carboxyl and carbonyl groups, which make these surfaces more polar. After adsorption of EO 37 PO 56 EO 37, a limited reduction in water contact angle, typically by 10, was observed for PP, PET, and nylon. In the case of (highly hydrophilic) cellulose, the contact angle was reduced by only 2. In agreement with AFM data, the thickness of the EO 37 - PO 56 EO 37 adsorbed layer after drying was nm, for the various substrates under investigation. Under different drying conditions, little difference in water contact angle and thickness of the adsorbed polymer layer was observed. Therefore, there was an indication that after treatment with EO 37 PO 56 EO 37 lower surface hydrophobicity was produced and the adsorbed layer was robust, independent of the drying conditions. Larger reductions in WCA were observed after surface treatment with similar amphiphilic polymers bearing more hydrophilic, larger EO end groups (data not shown). Therefore, a proper balance in the length of the EO blocks relative to the PO blocks can be used to target given adsorption masses and resultant WCA. The stability of adsorbed EO 37 PO 56 EO 37 on the different polymer surfaces was further tested by exposing the different samples to high energy sonication. In these experiments, after adsorption of EO 37 PO 56 EO 37 and rinsing, the polymer surfaces were subjected to sonication in water during 5, 10, and 15 min. After drying with a gentle nitrogen jet, the respective water Langmuir 2010, 26(12), contact angle was measured. An increase in WCA, compared to that measured for similar samples that were not subject to sonication, could be attributed to the release or depletion of the adsorbed layer. However, the results from these experiments revealed that only a very limited change in WCA, maximum increase of 2, was observed in the case of PP, PET and nylon, regardless of the time of sonication. Therefore, the adsorbed layers of EO 37 PO 56 EO 37 on these surfaces were very stable. In contrast, in the case of cellulose surfaces, sonication treatment produced a change of WCA consistent with the release of the adsorbed layer since the WCA returned to that typical of the bare cellulose surfaces. Thus, it can be concluded that adsorbed EO 37 PO 56 EO 37 on cellulose did not withstand the high-energy treatment, likely because of a weaker affinity with this substrate. Conclusions The adsorbed mass and the dynamics of adsorption for triblock nonionic copolymer EO 37 PO 56 EO 37 were investigated by combining QCM and SPR experiments using polymeric surfaces with different polarity. A larger EO 37 PO 56 EO 37 adsorption was quantified on the more hydrophobic surfaces (PP, PET, and nylon), while limited adsorption occurred on cellulose. In case of hydrophilic, polar surfaces, adsorption was observed with an AFM liquid cell to occur in the form spherical micellar structures with dimensions equivalent to those in the bulk solution. In contrast, featureless adsorbed layer structures, likely from dangling EO brushes, were formed on the more hydrophobic surfaces. The kinetics of adsorption was found to follow a Langmuir model and adsorption plateau at the various solution concentrations was observed close to the CMC of the EO 37 PO 56 EO 37 polymer. QCM dissipation factors revealed that after rinsing with water the adsorbed EO 37 PO 56 EO 37 formed rigid layers on the different DOI: /la100156a 9573

10 Article polymer substrates. By combining the areal mass calculated from QCM and SPR data, the contribution of coupled water in the adsorbed layers was quantified. The WCAs of the hydrophobic substrates after adsorption of EO 37 PO 56 EO 37 changed to a limited extent. This was due to the fact that less than half the limiting coverage, as measured by surface tension isotherms, was reached and also due to the relatively small number of EO groups in the EO 37 PO 56 EO 37 molecule. The thicknesses of the adsorbed EO 37 PO 56 EO 37 and WCAs of the respective surfaces after drying Liu et al. and sonication revealed that the adsorbed layers on PP, PET, and nylon surfaces were robust and stable. Acknowledgment. This project was supported by the Nonwovens Cooperative Research Center Project number Supporting Information Available: Additional information as described in the text. This material is available free of charge via the Internet at DOI: /la100156a Langmuir 2010, 26(12),